mirror of
https://github.com/openjdk/jdk.git
synced 2026-05-22 03:17:54 +00:00
8244340: Handshake processing thread lacks yielding
Reviewed-by: pchilanomate, dholmes, dcubed
This commit is contained in:
parent
9a0463165d
commit
95b8e9eaa3
@ -63,6 +63,94 @@ public:
|
||||
bool is_direct() { return _is_direct; }
|
||||
};
|
||||
|
||||
// Performing handshakes requires a custom yielding strategy because without it
|
||||
// there is a clear performance regression vs plain spinning. We keep track of
|
||||
// when we last saw progress by looking at why each targeted thread has not yet
|
||||
// completed its handshake. After spinning for a while with no progress we will
|
||||
// yield, but as long as there is progress, we keep spinning. Thus we avoid
|
||||
// yielding when there is potential work to be done or the handshake is close
|
||||
// to being finished.
|
||||
class HandshakeSpinYield : public StackObj {
|
||||
private:
|
||||
jlong _start_time_ns;
|
||||
jlong _last_spin_start_ns;
|
||||
jlong _spin_time_ns;
|
||||
|
||||
int _result_count[2][HandshakeState::_number_states];
|
||||
int _prev_result_pos;
|
||||
|
||||
int prev_result_pos() { return _prev_result_pos & 0x1; }
|
||||
int current_result_pos() { return (_prev_result_pos + 1) & 0x1; }
|
||||
|
||||
void wait_raw(jlong now) {
|
||||
// We start with fine-grained nanosleeping until a millisecond has
|
||||
// passed, at which point we resort to plain naked_short_sleep.
|
||||
if (now - _start_time_ns < NANOSECS_PER_MILLISEC) {
|
||||
os::naked_short_nanosleep(10 * (NANOUNITS / MICROUNITS));
|
||||
} else {
|
||||
os::naked_short_sleep(1);
|
||||
}
|
||||
}
|
||||
|
||||
void wait_blocked(JavaThread* self, jlong now) {
|
||||
ThreadBlockInVM tbivm(self);
|
||||
wait_raw(now);
|
||||
}
|
||||
|
||||
bool state_changed() {
|
||||
for (int i = 0; i < HandshakeState::_number_states; i++) {
|
||||
if (_result_count[0][i] != _result_count[1][i]) {
|
||||
return true;
|
||||
}
|
||||
}
|
||||
return false;
|
||||
}
|
||||
|
||||
void reset_state() {
|
||||
_prev_result_pos++;
|
||||
for (int i = 0; i < HandshakeState::_number_states; i++) {
|
||||
_result_count[current_result_pos()][i] = 0;
|
||||
}
|
||||
}
|
||||
|
||||
public:
|
||||
HandshakeSpinYield(jlong start_time) :
|
||||
_start_time_ns(start_time), _last_spin_start_ns(start_time),
|
||||
_spin_time_ns(0), _result_count(), _prev_result_pos(0) {
|
||||
|
||||
const jlong max_spin_time_ns = 100 /* us */ * (NANOUNITS / MICROUNITS);
|
||||
int free_cpus = os::active_processor_count() - 1;
|
||||
_spin_time_ns = (5 /* us */ * (NANOUNITS / MICROUNITS)) * free_cpus; // zero on UP
|
||||
_spin_time_ns = _spin_time_ns > max_spin_time_ns ? max_spin_time_ns : _spin_time_ns;
|
||||
}
|
||||
|
||||
void add_result(HandshakeState::ProcessResult pr) {
|
||||
_result_count[current_result_pos()][pr]++;
|
||||
}
|
||||
|
||||
void process() {
|
||||
jlong now = os::javaTimeNanos();
|
||||
if (state_changed()) {
|
||||
reset_state();
|
||||
// We spin for x amount of time since last state change.
|
||||
_last_spin_start_ns = now;
|
||||
return;
|
||||
}
|
||||
jlong wait_target = _last_spin_start_ns + _spin_time_ns;
|
||||
if (wait_target < now) {
|
||||
// On UP this is always true.
|
||||
Thread* self = Thread::current();
|
||||
if (self->is_Java_thread()) {
|
||||
wait_blocked((JavaThread*)self, now);
|
||||
} else {
|
||||
wait_raw(now);
|
||||
}
|
||||
_last_spin_start_ns = os::javaTimeNanos();
|
||||
}
|
||||
reset_state();
|
||||
}
|
||||
};
|
||||
|
||||
class VM_Handshake: public VM_Operation {
|
||||
const jlong _handshake_timeout;
|
||||
public:
|
||||
@ -81,7 +169,7 @@ class VM_Handshake: public VM_Operation {
|
||||
bool VM_Handshake::handshake_has_timed_out(jlong start_time) {
|
||||
// Check if handshake operation has timed out
|
||||
if (_handshake_timeout > 0) {
|
||||
return os::elapsed_counter() >= (start_time + _handshake_timeout);
|
||||
return os::javaTimeNanos() >= (start_time + _handshake_timeout);
|
||||
}
|
||||
return false;
|
||||
}
|
||||
@ -117,10 +205,7 @@ class VM_HandshakeOneThread: public VM_Handshake {
|
||||
VM_Handshake(op), _target(target) {}
|
||||
|
||||
void doit() {
|
||||
jlong start_time_ns = 0;
|
||||
if (log_is_enabled(Info, handshake)) {
|
||||
start_time_ns = os::javaTimeNanos();
|
||||
}
|
||||
jlong start_time_ns = os::javaTimeNanos();
|
||||
|
||||
ThreadsListHandle tlh;
|
||||
if (tlh.includes(_target)) {
|
||||
@ -131,13 +216,15 @@ class VM_HandshakeOneThread: public VM_Handshake {
|
||||
}
|
||||
|
||||
log_trace(handshake)("JavaThread " INTPTR_FORMAT " signaled, begin attempt to process by VMThtread", p2i(_target));
|
||||
jlong timeout_start_time = os::elapsed_counter();
|
||||
bool by_vm_thread = false;
|
||||
HandshakeState::ProcessResult pr = HandshakeState::_no_operation;
|
||||
HandshakeSpinYield hsy(start_time_ns);
|
||||
do {
|
||||
if (handshake_has_timed_out(timeout_start_time)) {
|
||||
if (handshake_has_timed_out(start_time_ns)) {
|
||||
handle_timeout();
|
||||
}
|
||||
by_vm_thread = _target->handshake_try_process(_op);
|
||||
pr = _target->handshake_try_process(_op);
|
||||
hsy.add_result(pr);
|
||||
hsy.process();
|
||||
} while (!_op->is_completed());
|
||||
|
||||
// This pairs up with the release store in do_handshake(). It prevents future
|
||||
@ -146,7 +233,7 @@ class VM_HandshakeOneThread: public VM_Handshake {
|
||||
// by the Handshakee.
|
||||
OrderAccess::acquire();
|
||||
|
||||
log_handshake_info(start_time_ns, _op->name(), 1, by_vm_thread ? 1 : 0);
|
||||
log_handshake_info(start_time_ns, _op->name(), 1, (pr == HandshakeState::_success) ? 1 : 0);
|
||||
}
|
||||
|
||||
VMOp_Type type() const { return VMOp_HandshakeOneThread; }
|
||||
@ -159,10 +246,7 @@ class VM_HandshakeAllThreads: public VM_Handshake {
|
||||
VM_HandshakeAllThreads(HandshakeOperation* op) : VM_Handshake(op) {}
|
||||
|
||||
void doit() {
|
||||
jlong start_time_ns = 0;
|
||||
if (log_is_enabled(Info, handshake)) {
|
||||
start_time_ns = os::javaTimeNanos();
|
||||
}
|
||||
jlong start_time_ns = os::javaTimeNanos();
|
||||
int handshake_executed_by_vm_thread = 0;
|
||||
|
||||
JavaThreadIteratorWithHandle jtiwh;
|
||||
@ -180,10 +264,10 @@ class VM_HandshakeAllThreads: public VM_Handshake {
|
||||
_op->add_target_count(number_of_threads_issued - 1);
|
||||
|
||||
log_trace(handshake)("Threads signaled, begin processing blocked threads by VMThread");
|
||||
const jlong start_time = os::elapsed_counter();
|
||||
HandshakeSpinYield hsy(start_time_ns);
|
||||
do {
|
||||
// Check if handshake operation has timed out
|
||||
if (handshake_has_timed_out(start_time)) {
|
||||
if (handshake_has_timed_out(start_time_ns)) {
|
||||
handle_timeout();
|
||||
}
|
||||
|
||||
@ -194,10 +278,13 @@ class VM_HandshakeAllThreads: public VM_Handshake {
|
||||
for (JavaThread *thr = jtiwh.next(); thr != NULL; thr = jtiwh.next()) {
|
||||
// A new thread on the ThreadsList will not have an operation,
|
||||
// hence it is skipped in handshake_try_process.
|
||||
if (thr->handshake_try_process(_op)) {
|
||||
HandshakeState::ProcessResult pr = thr->handshake_try_process(_op);
|
||||
if (pr == HandshakeState::_success) {
|
||||
handshake_executed_by_vm_thread++;
|
||||
}
|
||||
hsy.add_result(pr);
|
||||
}
|
||||
hsy.process();
|
||||
} while (!_op->is_completed());
|
||||
|
||||
// This pairs up with the release store in do_handshake(). It prevents future
|
||||
@ -257,10 +344,7 @@ bool Handshake::execute_direct(HandshakeClosure* thread_cl, JavaThread* target)
|
||||
JavaThread* self = JavaThread::current();
|
||||
HandshakeOperation op(thread_cl, /*is_direct*/ true);
|
||||
|
||||
jlong start_time_ns = 0;
|
||||
if (log_is_enabled(Info, handshake)) {
|
||||
start_time_ns = os::javaTimeNanos();
|
||||
}
|
||||
jlong start_time_ns = os::javaTimeNanos();
|
||||
|
||||
ThreadsListHandle tlh;
|
||||
if (tlh.includes(target)) {
|
||||
@ -270,14 +354,17 @@ bool Handshake::execute_direct(HandshakeClosure* thread_cl, JavaThread* target)
|
||||
return false;
|
||||
}
|
||||
|
||||
bool by_handshaker = false;
|
||||
HandshakeState::ProcessResult pr = HandshakeState::_no_operation;
|
||||
HandshakeSpinYield hsy(start_time_ns);
|
||||
while (!op.is_completed()) {
|
||||
by_handshaker = target->handshake_try_process(&op);
|
||||
HandshakeState::ProcessResult pr = target->handshake_try_process(&op);
|
||||
hsy.add_result(pr);
|
||||
// Check for pending handshakes to avoid possible deadlocks where our
|
||||
// target is trying to handshake us.
|
||||
if (SafepointMechanism::should_block(self)) {
|
||||
ThreadBlockInVM tbivm(self);
|
||||
}
|
||||
hsy.process();
|
||||
}
|
||||
|
||||
// This pairs up with the release store in do_handshake(). It prevents future
|
||||
@ -286,7 +373,7 @@ bool Handshake::execute_direct(HandshakeClosure* thread_cl, JavaThread* target)
|
||||
// by the Handshakee.
|
||||
OrderAccess::acquire();
|
||||
|
||||
log_handshake_info(start_time_ns, op.name(), 1, by_handshaker ? 1 : 0);
|
||||
log_handshake_info(start_time_ns, op.name(), 1, (pr == HandshakeState::_success) ? 1 : 0);
|
||||
|
||||
return op.executed();
|
||||
}
|
||||
@ -393,22 +480,22 @@ bool HandshakeState::claim_handshake(bool is_direct) {
|
||||
return false;
|
||||
}
|
||||
|
||||
bool HandshakeState::try_process(HandshakeOperation* op) {
|
||||
HandshakeState::ProcessResult HandshakeState::try_process(HandshakeOperation* op) {
|
||||
bool is_direct = op->is_direct();
|
||||
|
||||
if (!has_specific_operation(is_direct)){
|
||||
// JT has already cleared its handshake
|
||||
return false;
|
||||
return _no_operation;
|
||||
}
|
||||
|
||||
if (!possibly_can_process_handshake()) {
|
||||
// JT is observed in an unsafe state, it must notice the handshake itself
|
||||
return false;
|
||||
return _not_safe;
|
||||
}
|
||||
|
||||
// Claim the semaphore if there still an operation to be executed.
|
||||
if (!claim_handshake(is_direct)) {
|
||||
return false;
|
||||
return _state_busy;
|
||||
}
|
||||
|
||||
// Check if the handshake operation is the same as the one we meant to execute. The
|
||||
@ -416,13 +503,13 @@ bool HandshakeState::try_process(HandshakeOperation* op) {
|
||||
// by another JavaThread might be in progress.
|
||||
if (is_direct && op != _operation_direct) {
|
||||
_processing_sem.signal();
|
||||
return false;
|
||||
return _no_operation;
|
||||
}
|
||||
|
||||
// If we own the semaphore at this point and while owning the semaphore
|
||||
// can observe a safe state the thread cannot possibly continue without
|
||||
// getting caught by the semaphore.
|
||||
bool executed = false;
|
||||
ProcessResult pr = _not_safe;
|
||||
if (can_process_handshake()) {
|
||||
guarantee(!_processing_sem.trywait(), "we should already own the semaphore");
|
||||
log_trace(handshake)("Processing handshake by %s", Thread::current()->is_VM_thread() ? "VMThread" : "Handshaker");
|
||||
@ -431,11 +518,11 @@ bool HandshakeState::try_process(HandshakeOperation* op) {
|
||||
DEBUG_ONLY(_active_handshaker = NULL;)
|
||||
// Disarm after we have executed the operation.
|
||||
clear_handshake(is_direct);
|
||||
executed = true;
|
||||
pr = _success;
|
||||
}
|
||||
|
||||
// Release the thread
|
||||
_processing_sem.signal();
|
||||
|
||||
return executed;
|
||||
return pr;
|
||||
}
|
||||
|
||||
@ -96,7 +96,15 @@ public:
|
||||
process_self_inner();
|
||||
}
|
||||
}
|
||||
bool try_process(HandshakeOperation* op);
|
||||
|
||||
enum ProcessResult {
|
||||
_no_operation = 0,
|
||||
_not_safe,
|
||||
_state_busy,
|
||||
_success,
|
||||
_number_states
|
||||
};
|
||||
ProcessResult try_process(HandshakeOperation* op);
|
||||
|
||||
#ifdef ASSERT
|
||||
Thread* _active_handshaker;
|
||||
|
||||
@ -1360,7 +1360,7 @@ class JavaThread: public Thread {
|
||||
_handshake.process_by_self();
|
||||
}
|
||||
|
||||
bool handshake_try_process(HandshakeOperation* op) {
|
||||
HandshakeState::ProcessResult handshake_try_process(HandshakeOperation* op) {
|
||||
return _handshake.try_process(op);
|
||||
}
|
||||
|
||||
|
||||
Loading…
x
Reference in New Issue
Block a user